Provide an alternative means of counting mementos without requesting the TimeMap

[machawk1]

Most archives like the Internet Archive Wayback (IA) Machine has a user-facing page to display the distribution of their captures. An example is https://web.archive.org/web/20250000000000*/cnn.com for cnn.com.

On this page you will see some text that says something like, “Saved 965,319 times”. This is useful information that is returned quickly as compared to waiting for IA to respond to a request to the Memento API. You will notice that a request for the TimeMap for cnn.com takes awhile to get a response.

Mink displays this count as you browse, based on its response from the server-side aggregator, which minimally takes the amount of time of the slowest archive to respond to the aggregator’s request plus the overhead for the server-side aggregator to aggregate the responses. We should be able to optimize this based on the sum of the account from the various archives then only request the TimeMaps from the archives upon request by the user instead of by-default.